Agent sandboxes / Mosaic Sandbox alternative

A Mosaic Sandbox alternative for a machine that grows across tasks

Mosaic already preserves machine memory and disk through hibernation, snapshots, and its retained-identity archive path. Its shared object-storage volumes are a different persistence surface. Compare those layers and their retention boundaries with an agent whose system filesystem keeps growing.

Start buildingCompare capabilities ↓

Reviewed 2026-10-18 · Published by the Sandbox0 team

When Mosaic Sandbox fits

Mosaic fits CPU agent execution that benefits from its task/repository workflow, durable processes, Linux microVMs, SSH, live forks, and included RBAC/SSO/audit features. Its separate object-backed volumes can share datasets and outputs across machines.

When to consider Sandbox0

Consider Sandbox0 for a regional-object-storage-backed complete writable RootFS with automatic online growth and supported paused file operations. Preserve any required task orchestration, user/group tooling, desktop, or full-kernel behavior separately.

Capabilities

Mosaic Sandbox and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeMosaic SandboxNot established in this reviewSandbox0ttl and hard_ttl default to 0 (disabled). Configure idle pause and deletion policies for the same persistent sandbox identity; an idle running sandbox still accrues compute charges.Read documentation →
Persistent filesMosaic SandboxWhole-machine snapshots/archives retain disk; volumes retain separate shared files after machine deletion.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSMosaic SandboxThe archive retains machine memory and disk, distinct from the separate /mnt/data object-backed mount.View evidence ↗Sandbox0Persistence extends beyond /workspace to paths such as /usr/local, /etc, and /home. Runtime tmpfs and virtual filesystems are excluded.Read documentation →
Release compute, keep stateMosaic SandboxHibernation and archive resume retain state; a non-persistent TTL instead destroys the sandbox.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateMosaic SandboxMachine archives and live forks explicitly capture memory and disk; explicit pause can refuse busy workloads.View evidence ↗Sandbox0Experimental memory checkpoints can preserve compatible process state. Use memory: true on both pause and resume on a compatible deployment. A filesystem checkpoint alone does not preserve running processes.Read documentation →

Online RootFS growth

On growth-enabled images, Sandbox0 automatically expands the writable RootFS as free space runs low, within a configured sparse logical ceiling up to 64 TiB. Agent processes keep running. Growth is enabled in the qualified Cloud region; confirm activation for other regions. Existing images and snapshots retain their captured storage geometry. The ceiling is logical capacity, not preallocated host disk.

Machine snapshots retain system state, and shared volumes retain objects; neither reviewed guide establishes automatic online growth of the complete machine RootFS. Sandbox0’s enabled RootFS growth expands one filesystem while its agent keeps running.

Files while compute is paused

Sandbox0's Files API reads persistent files without resuming compute. Filesystem-only pauses also permit edits; retained-memory pauses are read-only and reject mutations with 409. SSH, SFTP, file watching, and runtime-only mounts require running compute.

Mosaic explicitly says files requests wake a hibernated sandbox. Its file API also restricts accessible roots; paths such as /etc require exec. Compare that with Sandbox0’s supported persistent-file access while compute remains paused.

See paused file operations →

Hibernation and an expiring archive are different states

Quiet guests can hibernate and resume on the next exec, SSH, or files request. At the separate TTL deadline, ordinary sandboxes are destroyed; persist=true instead archives memory and disk under the same sandbox ID. The guide reclaims an archive after seven days without resume. A retained ID does not eliminate retention policy.

Source 1 ↗Source 2 ↗

Shared object storage is not a general-purpose RootFS

A named volume can outlive sandboxes and be shared across hosts, while a machine snapshot retains its prepared tools. The volume guide documents object-storage latency and restrictions on reopening large existing objects for append or non-truncating writes. Put mutable dependency/build trees on the appropriate machine layer instead of assuming the shared mount has ordinary local-disk behavior.

Source 1 ↗Source 2 ↗

A billing unit is not a universal per-machine price

The public rate card charges one unit for the default 2 vCPU/4 GB shape; the smaller 1 vCPU shape still costs one unit, and 4 vCPU costs two. The current index also lists larger base-template shapes and directs callers to effective template limits. Preserve that size scope rather than price every shape at the headline rate.

Source 1 ↗Source 2 ↗

Pricing

Compare billing models, then price your workload

CPU allocation, memory, billing minimums, disk, plan fees, and supported workload differ. These meters are not equal-performance machine quotes. Confirm region, resource units, and included allowances before estimating a total.

Mosaic Sandbox

Pricing reviewed 2026-10-18

One active billing unit
$0.05 per hour
Default 2 vCPU/4 GB; 1 vCPU/2 GB also occupies one unit; 4 vCPU/8 GB occupies two.
Rate source ↗
Hibernated snapshot
$0
The public offer describes local NVMe snapshots with no idle charge.
Rate source ↗
Network egress
$0 per GB
Rate source ↗
Included active resource scope
RAM and storage included in the advertised compute offer
Confirm separate archive/volume retention and effective account tariffs.
Rate source ↗

Hibernated sandboxes count toward the advertised 200-alive-sandboxes organization limit. Public pricing and the current larger-shape catalog have different coverage; inspect the effective template limits and tariff before quoting those shapes. No equal-performance or storage-durability equivalence is inferred from the all-in unit price.

Check current Mosaic Sandbox pricing ↗

Sandbox0

$0.015/GiB-hour compute (billed by the second at approximately $0.000004167/GiB-second), plus $0.02/GiB-month of retained incremental RootFS and memory checkpoint data. Shared base images and platform recovery retention are excluded; paused sandboxes accrue storage charges only.

Compute / GiB-hour
$0.015
Retained data / GiB-month
$0.02
Paused compute
$0.00

A 4 GiB agent, 100 running hours

$6.00 compute

4 GiB × 100 hours, billed by the second. Keeping 20 GiB of billable incremental data for a full billing month adds $0.40. CPU is derived from memory by the platform; choose a tier that fits your workload.

Understand storage billing →

Before migrating from Mosaic Sandbox

  1. Inventory machine snapshots, persisted-ID archives and deadlines, named volumes, actual template shapes, durable processes, and any task/fork API dependencies.
  2. Export required machine and shared-volume data, then port compatible tools, execution, credentials, networking, and surrounding orchestration.
  3. Validate whole-system changes, online RootFS growth, paused file operations, retention/recovery behavior, and the effective size and storage tariff.

Common questions

Is Sandbox0 a drop-in Mosaic Sandbox replacement?

Evaluate API compatibility and product scope. A shared sandbox category does not make SDKs, lifecycle semantics, or limits interchangeable.

Does keeping files also keep a running agent?

Filesystem state and volatile process state are separate. Sandbox0 requires explicit compatible memory checkpoints to preserve running processes; its default filesystem resume starts fresh processes.

Sources and review scope

The Boat team's Battleships directory is the discovery source. References below describe the product and plan discussed on this page. Check current availability, regional terms and billing units before migrating. Where the cited material does not establish a capability or price, the comparison identifies that uncertainty.

Battleships research record ↗sandbox.mosaicos.com/docs/ ↗sandbox.mosaicos.com/llms.txt ↗sandbox.mosaicos.com/pricing ↗sandbox.mosaicos.com/ ↗Sandbox0 RootFS capacity →

Related comparisons

Agent 37 vs Sandbox0 →AgentComputer vs Sandbox0 →ainclave vs Sandbox0 →Alibaba Cloud Agent Sandbox / FC / AgentRun vs Sandbox0 →Alibaba Cloud AgentBay vs Sandbox0 →Amika vs Sandbox0 →